Performance Evaluation of a Voice-Based Depression Assessment System Considering the Number and Type of Input Utterances.

Takeshi Takano,Masakazu Higuchi,Kenji Miyazaki,Shuji Shinohara,Yasuhiro Omiya,Noriaki Sonota,Shinichi Tokuno,Shunji Mitsuyoshi,Mitsuteru Nakamura

doi:10.3390/s22010067

Takeshi Takano, Masakazu Higuchi + Show 7 more

Open Access

https://doi.org/10.3390/s22010067

Copy DOI

Abstract

It is empirically known that mood changes affect facial expressions and voices. In this study, the authors have focused on the voice to develop a method for estimating depression in individuals from their voices. A short input voice is ideal for applying the proposed method to a wide range of applications. Therefore, we evaluated this method using multiple input utterances while assuming a unit utterance input. The experimental results revealed that depressive states could be estimated with sufficient accuracy using the smallest number of utterances when positive utterances were included in three to four input utterances.

Highlights

Mental health care has become increasingly important worldwide due to stressful conditions in current society
Since the recordings were conducted at multiple facilities, we evaluated whether any potential difference was caused by the recording location; no significant differences were found in the standard vitality of the healthy group
The healthy group/depressive group discrimination accuracy is important for vitality; we evaluated the differences in the combinations of utterances that gave the maximum area under the curve (AUC) value

Summary

Introduction

Mental health care has become increasingly important worldwide due to stressful conditions in current society. Stress has a negative effect on health and mood in daily life, and accumulating stress induces mental and behavioral disorders [1]. These disorders cause lifetime income loss and decreased labor productivity, thereby resulting in a substantial economic loss to society [2,3]. Used non-invasive methods include self-administered psychological tests, such as the Patient Health Questionnaire 9 (PHQ9), General Health Questionnaire (GHQ), and Beck Depression Inventory (BDI) [7,8,9] These self-administered psychological tests are relatively simple to implement, they cannot rule out the drawback of reporting biases [10] wherein conscious and subconscious decisions of the respondent result in specific information being selectively underestimated or overestimated.

Methods

Results

Discussion

Conclusion